Ritesh Chimoriya, S. Pandey, Sharyn Carnahan et al. (6 authors) • Semantic Scholar
Finding: Systematic review of 19 studies (2020–2024) finds structured peer-supported learning (PAL, TBL, PLTL) in online/hybrid higher education settings is associated with improved academic performance, engagement, motivation, and well-being, though evidence quality is weak and dominated by small, descriptive designs.
Why it matters: Relevant to learning sciences, team-based collaboration, and motivation in digital learning contexts, with implications for designing peer learning structures in HCI-adjacent educational technology.
Method: Narrative, theory-informed synthesis using MMAT quality appraisal across 19 studies; meta-analysis was precluded by heterogeneity.
Jargon: PAL = peer-assisted learning; NPT = near-peer teaching; TBL = team-based learning; PLTL = peer-led team learning; MMAT = Mixed Methods Appraisal Tool for quality assessment.
M. Póli, Fotini Maniou, I. Maniou et al. (5 authors) • Semantic Scholar
Finding: Co-design with socially vulnerable groups (elderly, disabled, children, low-income) requires deliberate redesign of participatory processes themselves, not just application of standard methods — otherwise participation becomes tokenistic; proposes a five-part framework covering recruitment, technique adaptation, power-sharing, facilitator role, and impact evaluation.
Why it matters: Directly relevant to HCI/UX design, participatory design methods, and inclusive design — useful for teaching human-centered design principles and the ethics of co-creation with underrepresented users.
Method: Conceptual and integrative literature review synthesizing participatory design, HCI, inclusive design, and research ethics literature.
Jargon: *Co-design* = design process where end-users are active creative partners, not just research subjects; *tokenistic involvement* = superficial inclusion that creates the appearance of participation without genuine power-sharing; *participatory impact evaluation* = assessing whether inclusion actually shaped outcomes.
Abdul Latif, M. Yusuf, Umi Widyastuti et al. (5 authors) • Semantic Scholar
Finding: Systematic review mapping behavioral finance and financial behavior research, identifying cognitive biases (overconfidence, self-control), financial literacy, and decision-making as core themes across 78 studies from Scopus.
Why it matters: Touches on behavioral science and cognitive biases relevant to decision-making research, but is narrowly focused on financial contexts rather than HCI, behavior change design, or learning sciences.
Method: Bibliometric analysis using VOSviewer and Biblioshiny plus systematic literature review (SLR) — the methodology itself may be of interest for conducting literature reviews.
Adrián Valledor, Francisco Calles, Josefa Gómez et al. (4 authors) • Semantic Scholar
Finding: Applies Human-Centred AI (HCAI) principles to interpreter training via Design Science Research, deriving five design principles (learner agency, transparency/calibrated trust, cognitive load protection, contextualized assistance, AI as support) and evaluating interaction design through heuristic evaluation with 42 professional interpreters.
Why it matters: Directly addresses HCI, human-AI interaction design, and learning sciences in a high-cognitive-demand domain, offering a transferable framework for designing AI-assisted educational technology that preserves human agency.
Method: Design Science Research (DSR) with technology acceptance survey (n=42 professionals) and formative heuristic evaluation of the resulting artifact.
Jargon: *Design Science Research (DSR)*: methodology focused on building and evaluating artifacts (systems, frameworks) as research contributions; *Calibrated trust*: user confidence in AI that accurately reflects the system's actual reliability; *HCAI*: Human-Centred AI design philosophy prioritizing human oversight and agency over automation.
Finding: A theoretically-driven scaffolding intervention for diverse biomedical research teams significantly improved knowledge integration, which in turn predicted downstream performance outcomes (grants, publications, journal impact), mediated by coordination effectiveness.
Why it matters: Directly relevant to team dynamics, collaboration, and collective intelligence—provides an empirically tested intervention design for knowledge-diverse teams tackling complex problems, with practical implications for how to structure early-stage collaborations.
Method: Field-based quasi-experimental design with 26 newly formed biomedical collaborations (127 scientists), comparing scaffolded vs. non-scaffolded teams on knowledge integration and downstream performance.
Jargon: *Scaffolding* = structured support provided early in a process to help teams build shared understanding; *grand challenges* = complex, ill-defined societal/scientific problems requiring interdisciplinary collaboration; *knowledge integration* = successfully combining expertise from different specialties into coherent collaborative work.
⚠️ Single study (not replicated) | Limited statistical detail in abstract
Sophie Elizabeth Klitgaard, Grace Wang, D. Fessler et al. (4 authors) • Semantic Scholar
Finding: Social companion AI (SCAI) appeals to users by reducing "relational complexity" — the cognitive and social demands of human relationships — but this reduction may cause relationship displacement, social deskilling, and diminished motivation for real human connection.
Why it matters: Directly relevant to AI's impact on human collaboration and social behavior, with implications for behavioral science (social demotivation, habit formation around AI use) and HCI (design ethics of companion AI, dark patterns, addiction by design).
Jargon: *Social deskilling* = loss of interpersonal competency from reduced practice; *Relational complexity* = the cognitive/emotional demands and risks inherent in navigating human relationships; *SCAI* = Social Companion AI (e.g., Replika).
Finding: Audio-visual congruence between a virtual tour guide's appearance (human-like vs. anime-like) and voice style (serious vs. cute) affects user acceptance, mediated by perceived competence or enjoyment depending on the combination.
Why it matters: Directly relevant to HCI design principles—specifically how sensory cue matching shapes user perception and acceptance of AI agents—with implications for designing conversational agents, avatars, and AI-assisted interfaces.
Method: Three experimental studies testing congruence effects, mediation (perceived competence/enjoyment), and moderation (destination type) using Dual-system Theory as framework.
Jargon: *Dual-system Theory*: cognitive framework distinguishing fast/affective (System 1) and slow/deliberative (System 2) processing; *audio-visual congruence*: alignment between what users see and hear from an agent.
Finding: This paper presents a five-part analytical framework (Task Objective, World Crisis, Character Relationship, Action Process, Result Feedback) for designing environmental quests in story-driven games that carry genuine emotional meaning rather than reducing restoration to repetitive tasks.
Why it matters: Directly relevant to games design and gamification — specifically how narrative structure and quest design can create meaningful engagement and behavior change, with implications for serious games and persuasive technology aimed at environmental attitudes.
Method: Design-based reflection on the authors' own indie RPG development (*A Patch of Blue*), using the five-part chain as an analytical framework — essentially a practitioner case study.
Jargon: *Quest-based narrative* — using structured story missions to deliver meaning beyond mechanical objectives; *RPG (Role-Playing Game)* — a game genre emphasizing character progression and narrative choice.
Finding: In human-AI teaming for air traffic control, cognitive skills and data literacy are prioritized over social and physical competencies, suggesting AI integration shifts the nature of required human expertise toward higher-order cognitive and analytical abilities.
Why it matters: Directly relevant to AI's impact on human collaboration and work, and raises questions about how to train and assess competencies in AI-augmented professional environments.
Method: Best-Worst Method (BWM), a multi-criteria decision-making technique applied to a small purposive sample (n=19) to rank competency priority weights.
Jargon: *BWM (Best-Worst Method)*: a structured comparison technique where respondents identify the best and worst criteria, then derive priority weights through pairwise comparisons — more efficient than full pairwise comparison matrices.
Finding: Otome game engagement (emotional and financial investment) is positively associated with identity development, real-life social interaction, and romantic expectations among young adult Filipinas, but does not significantly predict psychological well-being.
Why it matters: Touches on user engagement mechanisms, parasocial relationships, and in-game purchase motivation—relevant to gamification, persuasive design, and the psychology of user engagement in games.
Method: Quantitative correlational design with Spearman's rho and multiple linear regression across 118 participants recruited from online gaming communities.
Jargon: *Otome games*: romance-focused narrative games targeting female players; *Parasocial interaction*: one-sided emotional relationships users form with media characters; *Uses and Gratifications Theory*: framework explaining why people seek out specific media based on needs they fulfill.
Yan Sun, Yang Gao, Xue-Ying Wang et al. (8 authors) • Semantic Scholar
Finding: Self-efficacy is the only significant mediator of a blended physical activity intervention's effect on MVPA, with moderation effects varying by demographics (age, sex, education, relationship status).
Why it matters: Marginally relevant to behavior change design (habit formation, persuasive technology) and intervention mapping, but the domain is workplace health/exercise rather than HCI, learning, or digital behavior change design specifically.
Method: 24-week RCT (N=141) with mediation and moderation analyses across three conditions (blended, web-only, control).
⭐ Longitudinal design ⚠️ Single study (not replicated)